python - 如何在 Pandas 数据框中找到 5 分钟的间隔？

Question

我有看起来与此类似的 pandas 数据框:

                     TIMESTAMP  EVENT_COUNT
0          2014-07-23 04:28:23            1 
1          2014-07-23 04:28:24            1
2   2014-07-23 04:28:25.999000            4
3          2014-07-23 04:28:27            1
4   2014-07-23 04:28:28.999000            2
5          2014-07-23 04:28:30            1
6          2014-07-23 04:29:31            7
7          2014-07-23 04:29:33            1
8          2014-07-23 04:29:34            1
9          2014-07-23 04:29:36            1
10         2014-07-23 04:40:37            2
11         2014-07-23 04:40:39            1
12         2014-07-23 04:40:40            1
13         2014-07-23 04:40:42            1
14         2014-07-23 04:40:43            1
15  2014-07-23 04:40:44.999000            4
16         2014-07-23 04:41:46            1
17         2014-07-23 04:41:47            1
18         2014-07-23 04:41:49            1
19         2014-07-23 04:41:50            1
20         2014-07-23 04:50:52            9
21         2014-07-23 04:50:53            4
22         2014-07-23 04:50:55            6
23         2014-07-27 01:12:13            1

我的最终目标是能够找出超过 5 分钟的空白。所以，从上面，我会发现之间的差距:

2014-07-23 04:29:36 and 2014-07-23 04:40:37
2014-07-23 04:41:50 and 2014-07-23 04:50:52
2014-07-23 04:50:55 and 2014-07-27 01:12:13

少于 5 分钟的间隙不需要识别。因此以下内容不会被识别为“差距”。

2014-07-23 04:28:30 and 2014-07-23 04:29:31    (Only 61 seconds)
2014-07-23 04:40:37 and 2014-07-23 04:40:39    (Only 2 seconds)
2014-07-23 04:40:44.999000 and 2014-07-23 04:41:46 (Just over 61 seconds)

如何找到上面提到的差距？当我尝试此 answer 中提到的解决方案时，似乎什么都没有改变。我使用了以下命令:

df.reindex(pd.date_range(min(df['TIMESTAMP']), max(df['TIMESTAMP']), freq='5min')).fillna(0)

运行此命令后数据框看起来相同。

Answer 1

IIUC 只要 dtype 已经是 datetime64 那么你就可以使用 diff这将创建一个时间增量，然后调用属性 dt.seconds :

In [8]:
df['OVER 5 MINS'] = (df['TIMESTAMP'].diff()).dt.seconds > 300
df

Out[8]:
                    TIMESTAMP  EVENT_COUNT OVER 5 MINS
INDEX                                                 
0     2014-07-23 04:28:23.000            1       False
1     2014-07-23 04:28:24.000            1       False
2     2014-07-23 04:28:25.999            4       False
3     2014-07-23 04:28:27.000            1       False
4     2014-07-23 04:28:28.999            2       False
5     2014-07-23 04:28:30.000            1       False
6     2014-07-23 04:29:31.000            7       False
7     2014-07-23 04:29:33.000            1       False
8     2014-07-23 04:29:34.000            1       False
9     2014-07-23 04:29:36.000            1       False
10    2014-07-23 04:40:37.000            2        True
11    2014-07-23 04:40:39.000            1       False
12    2014-07-23 04:40:40.000            1       False
13    2014-07-23 04:40:42.000            1       False
14    2014-07-23 04:40:43.000            1       False
15    2014-07-23 04:40:44.999            4       False
16    2014-07-23 04:41:46.000            1       False
17    2014-07-23 04:41:47.000            1       False
18    2014-07-23 04:41:49.000            1       False
19    2014-07-23 04:41:50.000            1       False
20    2014-07-23 04:50:52.000            9        True
21    2014-07-23 04:50:53.000            4       False
22    2014-07-23 04:50:55.000            6       False
23    2014-07-27 01:12:13.000            1        True